Data Science is among some of the newest and hottest ways that businesses are looking to exploit their data and meet their business objectives. The premise is intriguing: imagine using an AI, trained to understand your business, that can automatically provide prompts and alerts to customer service representatives to identify new up-sell and cross-sell opportunities as the conversation progresses. Alternatively, perhaps that same AI can guide a representative through a difficult complaint call to maximise customer satisfaction and retention.
The challenge, however, lies in the name: Data Science. The science element dictates its exploratory nature, looking to prove or disprove a hypothesis as to what data and algorithms may or may not be able to predict, automate, or visualise without necessarily having a clear view of what it will deliver from the outset. At the same time, however, a business understandably cannot be expected to pour money into a venture and just hope that sometime down the line it will repay that investment with interest. The leap of faith is something that all business owners will eventually need to take, but it is probably not a good idea to make it a cornerstone of your strategy.
The key is to see Data Science as a capability rather than as an asset for generating value. Time and investment need to be made to ensure that Data Science is embedded into your data strategy, your technical infrastructure and your analytics activities. Is your business agile enough to understand and support the iterative nature of Data Science activity? Do you have the technical infrastructure in place to ensure that the output of your research makes it into production and does not end up languishing in the proof-of-concept phase?
The technology world is changing rapidly, and the migration into cloud architecture at many enterprises is helping with some of these difficult questions. However, how many of these cloud deployments are being designed with Data Science capabilities in mind? Primarily, the investment in cloud is looking to rationalise infrastructure costs and provide an operational data storage solution for an ever-evolving Digital landscape. To overcome this, organisations require to look at their Data Strategy through the lens of their Technology investment case. The innovation departments, data departments, and technology must all be aligned if they are to all deliver on their potential. For ML or AI solutions to be successfully deployed, the data ingestion, ETL processes, and data pipelines need to be in place to allow the ML Ops teams to connect the solution the Data Science teams build. If the upfront investment in the right tech stack is justified based on the Data Science benefits it will enable and the data strategy is able to take into consideration the agile, iterative nature of research and development, then the Data Science use cases will stand a much better chance of success.
Favouring long term benefits, brought about through an investment in infrastructure that enables Data Science, over short-term use case benefits that will seldom be realised, should be at the forefront of any organisation’s Data Strategy. At Forecast we have a specialised team who can help our clients navigate the challenges and pitfalls of enabling a Data Science capability in their business. If you are interested in what you have read, then please do get in touch!
